Re: [ipwave] Shepherd review of draft-ietf-ipwave-ipv6-over-80211ocb-12 - Normative text in Ethernet Adaptation Layer

[Tony Li <[e-mail]>]

> > This presumes there is an associated Ethernet frame in the
> > picture somewhere.  What if there isn't? For example, V2V or V2I with no
> > Ethernet LAN involved.
> 
> Well, do you imply that some IP packet could be sent on something which is not 802.3?  Which is further not converted to an 802.11 frame?  I have never seen such thing, but I am curious what do you mean?


There are MANY historical media that are not Ethernet/802.3/802.11.  SLIP, PPP, SONET, ATM, etc….

802.3 is not a strict requirement, it’s just become the de-facto standard since just about everything else has aged out. 

Doing something else would require a truly compelling reason, and that’s not at all visible. Until then, let’s do the obvious thing, please.


> But I dont think the future use of 4 addresses in the 802.11 Data header will prevent mapping the Source to Source and the Destination to Destination (802.11 to Ethernet II headers).  Because if they do, that will break many other things, like RFC2464 and potentially others.
> 
> That will be a significant departure needing much agreement.


Concur.  Please stick with 2 addresses for now.

Tony